Sleep breathing sounds recorded via smartphone yielded fair prediction of obstructive sleep apnea, researchers reported in JAMA Otolaryngology – Head & Neck Surgery.

“The gold standard diagnostic method for OSA is attended, full-night, in-laboratory polysomnography that involves recording numerous physiologic signals that are manually scored by certified sleep technicians or physicians. Therefore, in-laboratory polysomnography is expensive, and accessibility to a sleep facility is not always easy,” Sung-Woo Cho, MD, from the department of otorhinolaryngology – head and neck surgery at Seoul National University Bundang Hospital at Seoul National University College of Medicine in Seongnam, South Korea, and colleagues wrote. “Considering the high prevalence of OSA, performing full-night, in-laboratory polysomnography may not be practical for all patients.”

The cross-sectional study enrolled 423 patients (mean age, 48.1 years; 84.1% men) who visited

the Seoul National University Bundang Hospital sleep center for snoring or sleep apnea from August 2015 to August 2019. Patients recorded audio during sleep with a smartphone during routine, full-night, in-laboratory polysomnography. Researchers conducted binary classifications for different threshold criteria based on an apnea hypopnea index threshold of five, 15 or 30 events per hour and created four regression models including noise reduction without feature selection, noise reduction with feature selection, neither noise reduction nor feature selection and feature selection without noise reduction.

Researchers split data into training (n = 256) and test (n = 167) data sets and patients were grouped as normal (n = 43), mild OSA (n= 80), moderate OSA (n = 109) or severe OSA (n = 191).

Smartphone-recorded audio yielded an accuracy of 88.2% for an apnea-hypopnea index threshold of five events per hour, 82.3% for 15 events per hour and 81.7% for 30 events per hour. Areas under the curve were 0.9, 0.89 and 0.9 for five, 15 and 30 events per hour, respectively.

All four regression models demonstrated similar results with correlation coefficients ranging from 0.77 to 0.78. Smartphone-recorded audio that was not denoised and had only selected attributes yielded the highest correlation coefficient in the regression analysis (r = 0.78). Both apnea-hypopnea index (beta = 0.33) and sleep efficiency (beta = –.2) were associated with OSA estimation error.

“These prediction models yielded a fair prediction performance, and we found that noise cancellation was not mandatory for a good prediction performance. … Future research should be extended to incorporate real-life smartphone recordings at home using various smartphone devices,” the researchers wrote.